Pat Gunn
Pat Gunn
@DrDaheeJang I'm grabbing it now; 32G is a bit big though; I'll test on this but it would've been fine to send a subset big enough to exhibit the issue...
@DrDaheeJang I have the data ready to go, but the link you provided doesn't provide me enough information; what notebook were you running? Could you make your modified notebook available?...
I started a run with your data after figuring out the right place to store the data for your script, and it ended up running for a long time, but...
I'll find a system with more RAM to run it locally and get back to you.
I reproduced the issue and am looking into it.
I think I see where in the code it's hitting the problem; I will need to talk with Kushal on how to fix it.
So I think what triggered the bug is that the (spatial, I think) background estimate has 3 components, and the third component is all zero. Two lines of inquiry: A)...
@DrDaheeJang In your opts_dict, try setting nb to 2 rather than 3 and see if it makes a difference. (even if this fixes it, it's still not great that the...
I spoke more with Kushal about this; having empty background components is something we should throw an exception over, telling people to reduce nb is the general right idea. I'm...
Sure. If you can send me another dataset as well as the script you're using with it (please send both - I really need the parameters), I can take a...